Information processing device, information processing method, and information processing program

Abstract

To reduce the burden of a user.SOLUTION: An information processing device 100 includes: an acquisition unit 131 that obtains text data input by a user; a generation unit 134 that generates a conversion candidate of the text data that satisfies a predetermined notation rule, by using a learned model that has learned the predetermined notation rule applied in a predetermined group, with an input of the text data obtained by the acquisition unit 131; and an output unit 135 that outputs the conversion candidate generated by the generation unit 134.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to an information processing device, an information processing method, and an information processing program. [Background technology]

[0002] Conventionally, when a user inputs characters, a function that predicts and presents a word that the user wants to input is known. For example, a predictive conversion technology using various language models is known (for example, Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Laid-Open No. 2014-178953 Summary of the Invention [Problem to be solved by the invention]

[0004] However, conventional technologies may not be able to reduce the burden on users. For example, conventional technologies predict characters entered by a user using data registered in a dictionary or the like and present them to the user as conversion candidates. However, if there are notation rules specific to an industry or company, there is a problem that conversion candidates that do not satisfy the notation rules are presented to the user. To address this, it is possible for the user to register conversion candidates that satisfy the notation rules, but this places a heavy burden on the user and may result in problems such as typos and omissions.

[0005] The present invention has been made in view of the above, and has an object to provide an information processing device, an information processing method, and an information processing program that reduce the burden on the user. [Means for solving the problem]

[0006] In order to solve the above-mentioned problems and achieve the object, the information processing device of the present invention is characterized by having an acquisition unit that acquires text data input by a user, a generation unit that uses the text data acquired by the acquisition unit as input and generates conversion candidates for the text data that satisfy the specified notation rules using a trained model that has learned specified notation rules to be applied within a specified group, and an output unit that outputs the conversion candidates generated by the generation unit. [Effects of the Invention]

[0007] According to the present invention, the burden on the user can be reduced. [Brief explanation of the drawings]

[0008] [Figure 1] FIG. 1 is a diagram illustrating an example of the configuration of an information processing system according to an embodiment. [Figure 2] FIG. 2 is a diagram illustrating an example of the configuration of the information processing device according to the embodiment. [Figure 3] FIG. 3 is a diagram illustrating an example of a format of a predetermined notation rule stored in a model storage unit according to the embodiment. [Figure 4] FIG. 4 is a diagram illustrating an example of a creation process performed by the prompt creation unit according to the embodiment. [Figure 5] FIG. 5 is a diagram for explaining an example of output processing by the output unit according to the embodiment. [Figure 6] FIG. 6 is a diagram illustrating an example of processing by the information processing device according to the embodiment. [Figure 7] FIG. 7 is a flowchart illustrating an example of the flow of processing by the information processing device according to the embodiment. [Figure 8] FIG. 8 is a diagram illustrating an example of a format of a predetermined notation rule stored in the model storage unit according to the embodiment. [Figure 9] FIG. 9 is a diagram illustrating an example of a creation process performed by the prompt creation unit according to the embodiment. [Figure 10]FIG. 10 is a diagram for explaining an example of output processing by the output unit according to the embodiment. [Figure 11] FIG. 11 is a diagram illustrating an example of an output process performed by the output unit according to the embodiment. [Figure 12] FIG. 12 is a diagram illustrating an example of an output process performed by the output unit according to the embodiment. [Figure 13] FIG. 13 is a diagram illustrating an example of an output process performed by the output unit according to the embodiment. [Figure 14] FIG. 14 is a diagram illustrating an example of processing by the information processing device according to the embodiment. [Figure 15] FIG. 15 is a flowchart illustrating an example of the flow of processing by the information processing device according to the embodiment. [Figure 16] FIG. 16 is a diagram illustrating an example of a computer that executes an information processing program. DETAILED DESCRIPTION OF THE INVENTION

[0009] Hereinafter, with reference to the drawings, an information processing device, an information processing method, and an information processing program according to the present application will be described in detail. Note that the present invention is not limited to these embodiments. In addition, in the description of the drawings, the same parts are denoted by the same reference numerals, and duplicated explanations will be omitted.

[0010] [Embodiment] [1. Information Processing System Configuration] An information processing system 1 shown in Fig. 1 will be described. Fig. 1 is a diagram showing an example of the configuration of the information processing system 1 according to the embodiment. As shown in Fig. 1, the information processing system 1 includes a terminal device 10 and an information processing device 100. The terminal device 10 and the information processing device 100 are connected to each other via a predetermined communication network (network N) so as to be able to communicate with each other via wired or wireless communication.

[0011] The terminal device 10 shown in FIG. 1 is a device such as a smartphone, a tablet terminal, a notebook PC, a desktop PC, a mobile phone, or a PDA.

[0012] The terminal device 10 is a device used for inputting data by a user. For example, the terminal device 10 transmits data input by a user to the information processing device 100. The terminal device 10 can also display, play, and the like data output from the information processing device 100.

[0013] 1 is a computer that can communicate with any terminal device 10 connected by wire or wirelessly. The information processing device 100 processes data acquired from the terminal device 10 and outputs the processed information to the terminal device 10.

[0014] The information processing device 100 is a computer intended to reduce the load on the user, and may be any device that can implement the processes in the embodiments.

[0015] For example, the information processing device 100 acquires text data input by a user, and uses the acquired text data as input to generate conversion candidates for the text data that satisfy the specified notation rules using a trained model that has learned specified notation rules that are applied within a specified group, and outputs the generated conversion candidates.

[0016] As a result, the information processing device 100 can reduce the burden on the user by generating conversion candidates that satisfy the notation rules for text data input by the user and displaying them on the terminal device 10. In other words, the information processing device 100 can reduce the effort required of the user when converting text data into text data that satisfies the notation rules by displaying conversion candidates that satisfy the notation rules on the terminal device 10, and also enables the user to observe the notation rules even if the user has forgotten them.

[0017] 2. Configuration of Information Processing Device Next, the configuration of the information processing device 100 will be described with reference to Fig. 2. Fig. 2 is a diagram showing an example of the configuration of the information processing device 100 according to the embodiment. As shown in Fig. 2, the information processing device 100 includes a communication unit 110, a control unit 130, and a storage unit 120. Note that these units may be held in a distributed manner in multiple devices. The processing of these units will be described below.

[0018] (Communication unit 110) The communication unit 110 is realized by a NIC (Network Interface Card) or the like, and enables communication between the control unit 130 and an external device via a telecommunication line such as a LAN or the Internet.

[0019] (Storage unit 120) The storage unit 120 is realized by a semiconductor memory element such as a RAM (Random Access Memory) or a flash memory, or a storage device such as a hard disk or an optical disk. The storage unit 120 has a user information storage unit 121, a model storage unit 122, an input information storage unit 123, and a generation information storage unit 124. Each unit of the storage unit 120 will be described below.

[0020] (User information storage unit 121) The user information storage unit 121 stores information about the user. For example, the user information storage unit 121 stores information such as the user's name, attributes, a setting for the speed at which conversion candidates are generated, and a setting for the accuracy of the conversion candidates.

[0021] (Model storage unit 122) The model storage unit 122 stores a trained model, predetermined notation rules, and prompts. An example of a trained model stored in the model storage unit 122 is a generation AI. An example of the generation AI is tsuzumi. tsuzumi is a natural language processing model fine-tuned to a specific field. Examples of specific fields include finance, medicine, semiconductors, IT (Information Technology), academia, factories (plants), law, and office services. tsuzumi was built with an emphasis on low power consumption and has a faster processing speed than ChatGPT. The model storage unit 122 stores multiple tsuzumis, each corresponding to a specific field such as finance, medicine, semiconductors, IT, academia, factories, law, or office services. The specific fields are not limited to the above.

[0022] An example of a generative AI is ChatGPT, a large-scale natural language processing model that is slower than tsuzumi but more accurate.

[0023] An example of the format of the predetermined notation rules stored in the model storage unit 122 will be described using Figure 3. Figure 3 is a diagram showing an example of the format of the predetermined notation rules stored in the model storage unit 122 according to the embodiment. "No." indicates the number of each rule included in the predetermined notation rules. "Recommended description" indicates the content of the recommended description. "Target description" indicates the content of the description that should be changed to the recommended description. For example, "1" in the predetermined notation rules stored in the model storage unit 122 indicates that when writing "server," the writing "server" is recommended.

[0024] (Input information storage unit 123) The input information storage unit 123 stores data input by the user, for example, text data input by the user.

[0025] (Generation information storage unit 124) The generation information storage unit 124 stores data generated by the generation unit 134 (described later). For example, the generation information storage unit 124 stores conversion candidates generated by the generation unit 134.

[0026] (control unit 130) The control unit 130 is realized using a CPU (Central Processing Unit), an NP (Network Processor), an FPGA (Field Programmable Gate Array), etc., and executes processing programs stored in memory. As shown in Fig. 2, the control unit 130 has an acquisition unit 131, a selection unit 132, a prompt creation unit 133, a generation unit 134, and an output unit 135. Each unit of the control unit 130 will be described below.

[0027] (Acquisition part 131) The acquiring unit 131 acquires text data input by the user. For example, the acquiring unit 131 acquires text data input by the user to the terminal device 10.

[0028] Furthermore, the acquiring unit 131 further acquires information about the material into which the text data is input, a setting of a speed for generating conversion candidates, and information about the accuracy of the conversion candidates. For example, the acquiring unit 131 acquires information about the material into which the text data is input, such as "field, content, and confidentiality (subject to disclosure)," by accepting input from the user.

[0029] Furthermore, the acquiring unit 131 acquires information on the setting of the speed at which conversion candidates are generated and the setting of the accuracy of the conversion candidates by referring to the user information storage unit 121. Furthermore, the acquiring unit 131 acquires information on the setting of the speed at which conversion candidates are generated and the setting of the accuracy of the conversion candidates by accepting input from the user.

[0030] (Selection unit 132) The selection unit 132 selects, as the trained model, either a first trained model which is a natural language processing model fine-tuned to a specific field, or a second trained model which is a large-scale natural language processing model, based on information about the material into which the text data acquired by the acquisition unit 131 is input, a setting for the speed at which conversion candidates are generated, and a setting for the accuracy of the conversion candidates.

[0031] More specifically, the selection unit 132 selects the first trained model (e.g., tsuzumi) when the material into which the text data is input belongs to a specific field (e.g., finance) and the speed at which the set conversion candidates are generated is faster than standard. Furthermore, the selection unit 132 selects the first trained model when the confidentiality level of the material into which the text data is input is equal to or higher than standard. When selecting the first trained model, the selection unit 132 selects the first trained model corresponding to the field to which the material into which the text data is input belongs.

[0032] On the other hand, when the material into which the text data is input does not belong to a specific field and the accuracy of the set conversion candidates is higher than the standard, the selection unit 132 selects the second trained model (for example, ChatGPT).Furthermore, when the confidentiality level of the material into which the text data is input is lower than the standard, the selection unit 132 selects the second trained model.

[0033] (Prompt creation unit 133) The prompt creation unit 133 creates a prompt that instructs the trained model to generate conversion candidates for text data that satisfy specified notation rules common to multiple materials shared within a specified group, and sets the created prompt in the trained model.

[0034] Here, an example of the creation process performed by the prompt creation unit 133 will be described with reference to FIG. 4. FIG. 4 is a diagram showing an example of the creation process performed by the prompt creation unit 133 according to the embodiment. FIG. 4(1) shows an example of a document shared by a predetermined group. FIG. 4(2) shows an example of a format in which predetermined notation rules are stored. "No." indicates the number of each rule included in the predetermined notation rules. "Recommended description" indicates the content of the recommended description. "Target description" indicates the content of the description that should be changed to the recommended description. FIG. 4(3) shows a model used by the generation unit 134.

[0035] For example, first, the prompt creation unit 133 extracts a predetermined notation rule "recommended description: server, target description: server, ..." that is common to proposal A and specification B from the data of proposal A and specification B shared within industry X, as shown in Figure 4(1) (Figure 4(2)). Note that the prompt creation unit 133 extracts the predetermined notation rule using existing natural language processing technology.

[0036] Next, the prompt creation unit 133 creates a prompt that instructs the trained model to generate conversion candidates for text data that satisfy the extracted notation rules “No. 1, recommended description: server, target description: server, No. 2...”, and sets the created prompt to the trained model (Figure 4(3)).

[0037] As another example, the prompt creation unit 133 can create a prompt that commands the generation of conversion candidates that satisfy the notation rules for each of a predetermined number of groups, and set the prompt in the trained model.

[0038] As an example, in the case where there are two groups and a specific notation rule common to the materials shared within each group exists, the prompt creation unit 133 creates a prompt that instructs the trained model to generate conversion candidates for text data that satisfy a first notation rule common to the multiple materials shared within the first group and a second notation rule common to the multiple materials shared within the second group, and sets the created prompt in the trained model.

[0039] More specifically, the prompt creation unit 133 creates a prompt that instructs the trained model to generate conversion candidates for text data that satisfy a first notation rule "recommended notation: server, target notation: server, ..." common to proposal A and specification B from the data of proposal A and specification B shared within industry X, and a second notation rule "recommended notation: person with a disability, target notation: person with a disability, ..." common to proposal C and proposal D from the data of proposal C and proposal D shared within company Y, and sets the created prompt in the trained model. Note that in the above example, for the sake of explanation, the number of groups and the number of notation rules are two, but the number of groups and notation rules is not limited to this.

[0040] (Generation unit 134) The generation unit 134 generates conversion candidates for text data. For example, the generation unit 134 receives the text data acquired by the acquisition unit 131 as input, and generates conversion candidates for text data that satisfy the predetermined notation rules using a trained model that has learned the predetermined notation rules that are applied within a predetermined group.

[0041] For example, the generation unit 134 receives the text data "server" acquired by the acquisition unit 131 as input and generates a conversion candidate for the text data "server" using a trained model that has learned a predetermined rule "recommended description: server, target description: server" that is applied within industry X.

[0042] That is, the generation unit 134 generates conversion candidates for text data that satisfy the predetermined notation rules by inputting the text data acquired by the acquisition unit 131 into a model that has learned predetermined notation rules that are applied within a predetermined group, such as a specific industry or company. Note that the generation unit 134 can generate conversion candidates using the first trained model or the second trained model selected by the selection unit 132.

[0043] (output unit 135) The output unit 135 outputs the conversion candidates generated by the generation unit 134. In addition, the output unit 135 may output conversion candidates derived by using existing technology in addition to the conversion candidates generated by the generation unit 134.

[0044] For example, when a user inputs text data "sa-ba" into a document, the conversion candidate "sa-ba" generated by the generation unit 134 and the conversion candidate "sa-ba" derived by using existing technology may be output together. In this case, when the user selects a conversion candidate different from the conversion candidate generated by the generation unit 134, the output unit 135 outputs a message to alert the user.

[0045] Here, an example of output processing by the output unit 135 will be described with reference to Fig. 5. Fig. 5 is a diagram showing an example of output processing by the output unit 135 according to the embodiment. Fig. 5 shows a document into which text data is input by a user. For example, when a conversion candidate "SERVER" different from the conversion candidate "SERVER" generated by the generation unit 134 is selected by the user, the output unit 135 displays on the terminal device 10, as shown in Fig. 5, a message to alert the user, "Does not meet company rules! Target description: SERVER Recommended description: SERVER."

[0046] [3. Processing] Next, an example of processing realized by the information processing device 100 according to the present embodiment will be described with reference to Fig. 6. Fig. 6 is a diagram illustrating an example of processing by the information processing device 100 according to the embodiment. Fig. 6 shows an example of executing processing to generate conversion candidates for acquired text data and output them to the terminal device 10.

[0047] The information processing device 100 acquires text data input by a user. The information processing device 100 creates and sets a prompt. For example, the information processing device 100 creates a prompt that instructs a trained model to generate conversion candidates for text data that satisfy predetermined notation rules common to multiple materials shared within a predetermined group, and sets the created prompt in the trained model.

[0048] The information processing device 100 receives text data as input and generates conversion candidates for the text data that satisfy the predetermined notation rules using a trained model that has learned the predetermined notation rules that are applied within a predetermined group.

[0049] Then, the information processing device 100 outputs the generated conversion candidates to the terminal device 10. For example, the information processing device 100 causes the terminal device 10 to display the generated conversion candidates.

[0050] [4. Flowchart] Next, the flow of processing by the information processing system 1 will be described with reference to Fig. 7. Note that the steps below may be executed in a different order, and some processing may be omitted.

[0051] First, the acquisition unit 131 determines whether or not an instruction to create conversion candidates has been received from the user (step S101). If the acquisition unit 131 determines that an instruction to create conversion candidates has not been received from the user (step S101: No), the process of step S101 is performed again.

[0052] On the other hand, if the acquisition unit 131 determines that it has received an instruction from the user to create conversion candidates (step S101: Yes), the acquisition unit 131 acquires the text data entered by the user, information about the material into which the text data is entered, a setting for the speed at which conversion candidates are generated, and a setting for the accuracy of the conversion candidates (step S102).

[0053] Next, the selection unit 132 selects a model (step S103). For example, the selection unit 132 selects, as the trained model, a first trained model which is a natural language processing model fine-tuned to a specific field, or a second trained model which is a large-scale natural language processing model, based on information about the material into which the text data acquired by the acquisition unit 131 is input, a setting for the speed at which conversion candidates are generated, and a setting for the accuracy of the conversion candidates.

[0054] Next, the prompt creation unit 133 creates a prompt (step S104). For example, the prompt creation unit 133 creates a prompt that instructs the trained model to generate conversion candidates for text data that satisfy predetermined notation rules common to multiple materials shared within a predetermined group.

[0055] Next, the prompt creation unit 133 sets the created prompt to the trained model (step S105). For example, the prompt creation unit 133 sets the created prompt to the trained model selected by the selection unit 132.

[0056] Next, the generation unit 134 generates conversion candidates for the text data (step S106). For example, the generation unit 134 receives the text data acquired by the acquisition unit 131 as input, and generates conversion candidates for the text data that satisfy the predetermined notation rules by using a trained model that has learned the predetermined notation rules that are applied within a predetermined group.

[0057] Next, the output unit 135 outputs the conversion candidates generated by the generation unit 134 (step S107). For example, the output unit 135 causes the terminal device 10 to display the conversion candidates generated by the generation unit 134. Then, the information processing device 100 ends the process.

[0058] [5. Effects] The information processing device 100 according to the embodiment includes an acquisition unit 131 that acquires text data input by a user, a generation unit 134 that uses the text data acquired by the acquisition unit 131 as input and generates conversion candidates for text data that satisfy the specified notation rules using a trained model that has learned the specified notation rules to be applied within a specified group, and an output unit 135 that outputs the conversion candidates generated by the generation unit 134.

[0059] As a result, the information processing device 100 can reduce the burden on the user by generating conversion candidates that satisfy the notation rules for text data input by the user and displaying them on the terminal device 10. In other words, the information processing device 100 can reduce the effort required of the user when converting text data into text data that satisfies the notation rules by displaying conversion candidates that satisfy the notation rules on the terminal device 10, and also enables the user to observe the notation rules even if the user has forgotten them.

[0060] In addition, the acquisition unit 131 of the information processing device 100 according to the embodiment further acquires information regarding the material into which text data is input, a setting for the speed at which conversion candidates are generated, and a setting for the accuracy of the conversion candidates, and a selection unit 132 selects, as the trained model, a first trained model which is a natural language processing model fine-tuned to a specific field, or a second trained model which is a large-scale natural language processing model, based on the information regarding the material into which text data is input acquired by the acquisition unit 131, the setting for the speed at which conversion candidates are generated, and the setting for the accuracy of the conversion candidates, and the generation unit 134 generates conversion candidates using the first trained model or the second trained model selected by the selection unit 132.

[0061] As a result, the information processing device 100 can reduce the burden on the user by generating conversion candidates that satisfy the notation rules for the text data entered by the user using a model selected according to the field, content, confidentiality of the material into which the text data is entered, settings for the speed at which conversion candidates are generated, and settings for the accuracy of the conversion candidates, and displaying these on the terminal device 10.

[0062] The information processing device 100 according to the embodiment has a prompt creation unit 133 that creates a prompt to instruct a trained model to generate conversion candidates for text data that satisfy predetermined notation rules common to multiple materials shared within a predetermined group, and sets the created prompt in the trained model.

[0063] As a result, the information processing device 100 can input text data into a model that has a prompt set to instruct the device to generate conversion candidates for text data that satisfy specified notation rules common to materials shared within a specified group, making it easier to comply with notation rules within a specified group, such as an industry or company, and reducing the burden on the user.

[0064] When a conversion candidate different from the conversion candidate generated by the generation unit 134 is selected by the user, the output unit 135 of the information processing device 100 according to the embodiment displays a message on the terminal device to alert the user.

[0065] As a result, when a user selects a conversion candidate that does not satisfy predetermined notation rules, the information processing device 100 can reduce the burden on the user by displaying a warning pop-up or the like on the terminal device 10.

[0066] [Variations] Up to this point, an example of generating conversion candidates for text data has been described. Below, a modified example will be described in which data of various formats is input and data of a different format corresponding to the input data is generated. Here, data of different formats may be data in which the data itself is saved in a different format, such as image data and text data, or data in which the written format is different, such as text data indicating a certain matter and text data indicating a source of content indicating a certain matter. In other words, the input data and the generated data may be saved in the same format. Content common to the embodiment will be omitted as appropriate.

[0067] 1. Configuration of Information Processing Device (User information storage unit 121) The user information storage unit 121 stores information about the user. For example, the user information storage unit 121 stores information such as the user's name, gender, age, setting of the speed at which data in different formats is generated, setting of the accuracy of data in different formats, and materials created by the user.

[0068] (Model storage unit 122) The model storage unit 122 stores a trained model, predetermined notation rules, and prompts. An example of the format of the predetermined notation rules stored in the model storage unit 122 will be described with reference to FIG. 8. FIG. 8 is a diagram showing an example of the format of the predetermined notation rules stored in the model storage unit 122 according to the embodiment. "No." indicates the number of each rule included in the predetermined notation rule. "Content" indicates the content of the notation rule. For example, "1" stored in the model storage unit 122 means a notation rule that uses the "desuu" style of writing. Furthermore, "2" stored in the model storage unit 122 means a notation rule that uses "a period at the end of a sentence."

[0069] (Input information storage unit 123) The input information storage unit 123 stores data input by the user, such as text data, table data, image data, and audio data input by the user.

[0070] (Generation information storage unit 124) The generated information storage unit 124 stores data generated by the generating unit 134, which will be described later. For example, the generated information storage unit 124 stores data in a different format corresponding to the input data generated by the generating unit 134.

[0071] (Acquisition part 131) The acquisition unit 131 acquires data that the user has input into the document. For example, the acquisition unit 131 acquires text data, table data, image data, audio data, etc. that the user has input into the document, for example, a Word file. Note that the format of the data acquired by the acquisition unit 131 does not matter.

[0072] Furthermore, the acquiring unit 131 further acquires information about the material into which the text data is input, a setting of the speed at which data in different formats is generated, and information about the accuracy of the data in different formats. For example, the acquiring unit 131 acquires information about the material into which the text data is input, such as "field, content, and confidentiality (subject to disclosure)," by accepting input from the user.

[0073] Furthermore, the acquiring unit 131 refers to the user information storage unit 121 to acquire information on the setting of the speed at which data in different formats is generated and the setting of the accuracy of data in different formats.

[0074] (Selection unit 132) The selection unit 132 selects, as the trained model, either a first trained model which is a natural language processing model fine-tuned to a specific field, or a second trained model which is a large-scale natural language processing model, based on information about the material into which the data acquired by the acquisition unit 131 is input, a setting for the speed at which data of different formats is generated, and a setting for the accuracy of data of different formats.

[0075] More specifically, the selection unit 132 selects the first trained model (e.g., tsuzumi) when the material into which the data is input belongs to a specific field (e.g., finance) and the speed for generating data in a different format is set faster than standard. Furthermore, the selection unit 132 selects the first trained model when the confidentiality level of the material into which the data is input is equal to or higher than standard. When selecting the first trained model, the selection unit 132 selects the first trained model corresponding to the field to which the material into which the data is input belongs.

[0076] On the other hand, the selection unit 132 selects the second trained model (e.g., ChatGPT) when the material into which the data is input does not belong to a specific field and the accuracy setting of the different format data is higher than the standard. Also, the selection unit 132 selects the second trained model when the confidentiality level of the material into which the data is input is lower than the standard.

[0077] (Prompt creation unit 133) The prompt creation unit 133 creates a prompt that instructs the trained model to generate data in a format different from the data corresponding to the data acquired by the acquisition unit 131, and sets the created prompt in the trained model. At this time, the prompt creation unit 133 may create a prompt that instructs the trained model to generate data in a specified format. For example, the prompt creation unit 133 accepts a specification of a different format, creates a prompt that instructs the trained model to generate data in the accepted format, and sets the created prompt in the trained model.

[0078] The prompt creation unit 133 further creates a prompt that instructs the trained model to create data that satisfies predetermined notation rules common to multiple materials created by the user, and sets the created prompt in the trained model. Here, an example of the creation process performed by the prompt creation unit 133 will be described with reference to FIG. 9. FIG. 9 is a diagram showing an example of the creation process performed by the prompt creation unit 133 according to the embodiment. FIG. 9(1) shows a material created by a user. FIG. 9(2) shows an example of a format in which predetermined notation rules are stored. "No." indicates the number of each rule included in the predetermined notation rules. "Content" indicates the content of the notation rule. FIG. 9(3) shows a model used by the generation unit 134.

[0079] For example, first, the prompt creation unit 133 extracts predetermined notation rules "No. 1: dearu-cho, No. 2: period at the end of the sentence, No. 3..." that are common to reports A and B created by user X, as shown in Figure 9(1) (Figure 9(2)). The prompt creation unit 133 extracts the predetermined notation rules using existing natural language processing technology.

[0080] Next, the prompt creation unit 133 creates a prompt that instructs the trained model to generate data that satisfies the extracted notation rules "No. 1: dearu tone, No. 2: period at the end of the sentence, No. 3...", and sets the created prompt in the trained model (Figure 9 (3)).

[0081] (Generation unit 134) The generation unit 134 receives the data acquired by the acquisition unit 131 as input, and generates data corresponding to the data in a format different from the data in question using a trained model that has learned the relationship between the data entered in the document and data in a format different from the data entered in the document that corresponds to the data entered in the document.

[0082] The generation unit 134 inputs the data acquired by the acquisition unit 131 into a model that has learned the relationship between certain data in the document and data in another format that corresponds to the certain data in the document, and generates data in a different format that corresponds to the acquired data. Note that the generation unit 134 can generate data in a different format using the first trained model or the second trained model selected by the selection unit 132.

[0083] Here, the generation process performed by the generation unit 134 will be described with reference to Figs. 10 to 13. Figs. 10 to 13 are diagrams showing an example of the generation process performed by the generation unit 134 according to the embodiment. Fig. 10 shows the correspondence between image data and text data. Fig. 11 shows the correspondence between table data and text data. Fig. 12 shows the correspondence between a flowchart and text data. Fig. 13 shows the correspondence between text data indicating a certain matter and text data indicating a citation source.

[0084] For example, when the data acquired by the acquisition unit 131 is image data (FIG. 10(1)), the generation unit 134 generates text data (FIG. 10(2)) that explains the acquired image data. Note that, contrary to the example described above, the generation unit 134 can also generate image data from text data that explains image data, the content of which is explained by the text data.

[0085] For example, when the data acquired by the acquisition unit 131 is table data (FIG. 11(1)), the generation unit 134 generates text data (FIG. 11(2)) that explains the acquired table data. Note that, contrary to the example described above, the generation unit 134 can also generate table data from text data that explains table data, the table data being the content explained by the text data.

[0086] For example, when the data acquired by the acquisition unit 131 is a flowchart (FIG. 12(1)), the generation unit 134 generates text data (FIG. 12(2)) that explains the acquired flowchart. Note that, contrary to the example described above, the generation unit 134 can also generate table data from the text data, the content of which is explained by the text data.

[0087] For example, when the data acquired by the acquisition unit 131 is text data requiring a quote source (FIG. 13(1)), the generation unit 134 generates text data indicating the quote source of the content indicated by the acquired text data (FIG. 13(2)). Note that, contrary to the above example, the generation unit 134 can also generate text data of the content indicated by the quote source from text data indicating the quote source.

[0088] Furthermore, the generation unit 134 can receive a specification of the format of the data to be generated and generate data in the specified format. For example, when the generation unit 134 receives a specification of the format from the user to generate data in a predetermined format, the generation unit 134 generates data in the specified format that corresponds to the data acquired by the acquisition unit 131.

[0089] (output unit 135) The output unit 135 outputs the data in a different format generated by the generation unit 134. For example, the output unit 135 outputs the data in a different format generated by the generation unit 134 to the terminal device 10. More specifically, when the data acquired by the acquisition unit 131 is image data and the generation unit 134 generates text data explaining the image data, the output unit 135 writes the generated text data in a location in the document corresponding to the location where the image data is pasted.

[0090] When the data acquired by the acquisition unit 131 is text data explaining the table data and the table data is generated by the generation unit 134, the output unit 135 pastes the generated table data in a location corresponding to the location in the document where the text data explaining the table data is written.

[0091] [2. Processing] Next, an example of processing realized by the information processing device 100 according to the present embodiment will be described with reference to Fig. 14. Fig. 14 is a diagram illustrating an example of processing by the information processing device 100 according to the embodiment. Fig. 14 shows an example of executing processing to generate conversion candidates for acquired text data and output them to the terminal device 10.

[0092] The information processing device 100 acquires data input into a document by a user. The information processing device 100 creates and sets a prompt. For example, the information processing device 100 creates a prompt that instructs a trained model to generate data in a different format corresponding to the data input into the document, and sets the created prompt in the trained model.

[0093] The information processing device 100 uses the acquired data as input and generates data of a different format corresponding to the acquired data using a trained model that has learned the relationship between the data entered in the document and data of a different format corresponding to the data entered in the document.

[0094] Then, the information processing device 100 outputs the generated data in a different format to the terminal device 10. For example, the information processing device 100 causes the terminal device 10 to display the generated data in a different format.

[0095] [3. Flowchart] Next, the flow of processing by the information processing system 1 will be described with reference to Fig. 15. Note that the steps below may be executed in a different order, and some processing may be omitted.

[0096] First, the acquiring unit 131 determines whether or not an instruction to create data in a different format has been received (step S201). If the acquiring unit 131 determines that an instruction to create data in a different format has not been received (step S201: No), the acquiring unit 131 performs the process of step S201 again.

[0097] On the other hand, if the acquisition unit 131 determines that it has received an instruction to create data in a different format (step S201: Yes), the acquisition unit 131 acquires the data entered by the user, information about the material into which the data is entered, settings for the speed at which the data is generated, and settings for the accuracy of the data (step S202).

[0098] Next, the selection unit 132 selects a model (step S203). For example, the selection unit 132 selects, as the trained model, a first trained model that is a natural language processing model fine-tuned to a specific field, or a second trained model that is a large-scale natural language processing model, based on information about the material into which the data acquired by the acquisition unit 131 is input, a setting for the speed at which the data is generated, and a setting for the accuracy of the data.

[0099] Next, the prompt creation unit 133 creates a prompt (step S204). For example, the prompt creation unit 133 creates a prompt that instructs the trained model to generate data corresponding to the data acquired by the acquisition unit 131 but in a format different from the data.

[0100] Next, the prompt creation unit 133 sets the created prompt to the trained model (step S205). For example, the prompt creation unit 133 sets the created prompt to the trained model selected by the selection unit 132.

[0101] Next, the generation unit 134 generates data in a different format corresponding to the data acquired by the acquisition unit 131 (step S206). For example, the generation unit 134 inputs the data acquired by the acquisition unit 131 into a model that has learned the relationship between certain data in the document and data in another format corresponding to the certain data in the document, and generates data in a different format corresponding to the acquired data.

[0102] Then, the output unit 135 outputs the data in the different format generated by the generation unit 134 (step S207). For example, the output unit 135 causes the terminal device 10 to display the data in the different format generated by the generation unit 134. Then, the information processing device 100 ends the process.

[0103] [4. Effects] The information processing device 100 according to the embodiment includes an acquisition unit 131 that acquires data entered into a document by a user, a generation unit 134 that uses the data acquired by the acquisition unit 131 as input and generates data in a format different from the data corresponding to the data using a trained model that has learned the relationship between the data entered into the document and data in a format different from the data entered into the document corresponding to the data entered into the document, and an output unit 135 that outputs the data in the format different from the data generated by the generation unit 134.

[0104] This allows the information processing device 100 to generate data in a different format corresponding to the data input by the user, thereby reducing the burden on the user when creating data such as explanatory text, tables, images, and citation sources.

[0105] In addition, the information processing device 100 according to the embodiment has a prompt creation unit 133 that creates a prompt to instruct the trained model to generate data in a format different from the data corresponding to the data acquired by the acquisition unit 131, and sets the created prompt in the trained model.

[0106] This allows the information processing device 100 to input data into a model that has a prompt set to instruct the device to generate data in a format different from the data corresponding to the acquired data, and generate data in a format different from the data entered by the user, thereby reducing the burden on the user when creating data such as descriptions, tables, images, and citation sources.

[0107] In addition, the prompt creation unit 133 of the information processing device 100 according to the embodiment further creates a prompt that instructs the trained model to generate data that satisfies predetermined notation rules common to multiple materials created by the user, and sets the created prompt in the trained model.

[0108] As a result, the information processing device 100 inputs data into a model that has a prompt set to instruct the device to generate data that satisfies specified notation rules common to multiple documents created by the user, and generates data in a different format that corresponds to the data entered by the user, thereby creating data such as explanatory text, tables, images, and citation sources that the user would create, thereby reducing the burden on the user.

[0109] [5. Program] It is also possible to create a program written in a computer-executable language that executes the processes executed by the information processing device 100 described in the above embodiment. In this case, the same effects as those of the above embodiment can be achieved by having a computer execute the program. Furthermore, such a program may be recorded on a computer-readable recording medium, and the program recorded on the recording medium may be read and executed by a computer to achieve the same processes as those of the above embodiment.

[0110] Fig. 16 is a diagram showing an example of a computer that executes an information processing program. As shown in Fig. 16, a computer 1000 includes, for example, a memory 1010, a CPU 1020, a hard disk drive interface 1030, a disk drive interface 1040, a serial port interface 1050, a video adapter 1060, and a network interface 1070. These components are connected by a bus 1080.

[0111] The memory 1010 includes a ROM (Read Only Memory) 1011 and a RAM 1012. The ROM 1011 stores, for example, a boot program such as a BIOS (Basic Input Output System). The hard disk drive interface 1030 is connected to a hard disk drive 1090. The disk drive interface 1040 is connected to a disk drive 1100. A removable storage medium such as a magnetic disk or optical disk is inserted into the disk drive 1100. The serial port interface 1050 is connected to, for example, a mouse 1110 and a keyboard 1120. The video adapter 1060 is connected to, for example, a display 1130.

[0112] 16, the hard disk drive 1090 stores, for example, an OS (Operating System) 1091, an application program 1092, a program module 1093, and program data 1094. The tables described in the above embodiments are stored in the hard disk drive 1090 or memory 1010, for example.

[0113] The information processing program is stored in the hard disk drive 1090 as a program module in which instructions to be executed by the computer 1000 are written. Specifically, the hard disk drive 1090 stores a program module 1093 in which each process executed by the computer 1000 described in the above embodiment is written.

[0114] Furthermore, data used for information processing by the information processing program is stored as program data, for example, in the hard disk drive 1090. Then, the CPU 1020 reads out the program module 1093 and program data 1094 stored in the hard disk drive 1090 into the RAM 1012 as necessary, and executes each of the above-described procedures.

[0115] The program module 1093 and program data 1094 related to the information processing program are not limited to being stored in the hard disk drive 1090, but may be stored in a removable storage medium and read by the CPU 1020 via the disk drive 1100, etc. Alternatively, the program module 1093 and program data 1094 related to the control program may be stored in another computer connected via a network such as a LAN or a WAN (Wide Area Network), and read by the CPU 1020 via the network interface 1070.

[0116] [7. Other] Although various embodiments have been described in detail herein with reference to the drawings, these embodiments are merely examples and are not intended to limit the present invention. The features described herein can be realized in various ways, including various modifications and improvements based on the knowledge of those skilled in the art.

[0117] Furthermore, the above-mentioned "module (-er suffix, -or suffix)" can be read as a unit, means, circuit, etc. For example, a communication module, a control module, and a storage module can be read as a communication unit, a control unit, and a storage unit, respectively. [Explanation of symbols]

[0118] 1. Information Processing Systems 100 Information processing device 110 Communications Department 120 Storage section 121 User information storage unit 122 Model Memory Unit 123 Input information storage unit 124 Generation information storage unit 130 control section 131 Acquisition Department 132 Selection section 133 Prompt Creation Department 134 Generation part 135 Output section

Claims

1. an acquisition unit that acquires text data input by a user; a generation unit that receives the text data acquired by the acquisition unit as an input, and generates conversion candidates for the text data that satisfy a predetermined notation rule using a trained model that has learned the predetermined notation rule to be applied within a predetermined group; an output unit that outputs the conversion candidates generated by the generation unit; An information processing device comprising:

2. The acquisition unit further acquires information about a document into which the text data is input, a setting of a speed for generating the conversion candidates, and a setting of accuracy of the conversion candidates, a selection unit that selects, as the trained model, a first trained model that is a natural language processing model fine-tuned to a specific field, or a second trained model that is a large-scale natural language processing model, based on information about a material into which the text data acquired by the acquisition unit is input, a setting for a speed at which the conversion candidates are generated, and a setting for accuracy of the conversion candidates; The generation unit generates the conversion candidates using the first trained model or the second trained model selected by the selection unit.

2. The information processing apparatus according to claim 1, wherein:

3. a prompt creation unit that creates a prompt that instructs the trained model to generate conversion candidates for text data that satisfy a predetermined notation rule common to a plurality of materials shared within the predetermined group, and sets the created prompt in the trained model; 2. The information processing apparatus according to claim 1, further comprising:

4. The output unit When the user selects a conversion candidate different from the conversion candidate generated by the generation unit, a message to alert the user is displayed on the terminal device.

2. The information processing apparatus according to claim 1, wherein:

5. an acquisition unit that acquires data entered by a user into the document; a generation unit that uses the data acquired by the acquisition unit as an input and generates data in a format different from the data corresponding to the data, using a trained model that has learned the relationship between data input to a document and data in a format different from the data input to the document corresponding to the data input to the document; an output unit that outputs the data in the different formats generated by the generation unit; An information processing device comprising:

6. a prompt creation unit that creates a prompt that instructs the trained model to generate data in a format different from the data corresponding to the data acquired by the acquisition unit, and sets the created prompt in the trained model.

6. The information processing apparatus according to claim 5, further comprising:

7. The prompt creation unit further creates a prompt that instructs the trained model to generate data that satisfies a predetermined notation rule common to a plurality of materials created by the user, and sets the created prompt in the trained model.

7. The information processing apparatus according to claim 6,

8. 1. A computer-implemented information processing method, comprising: an acquisition step of acquiring text data input by a user; a generation step of generating conversion candidates for the text data that satisfy a predetermined notation rule using a trained model that has learned a predetermined notation rule applied within a predetermined group, with the text data acquired by the acquisition step as an input; an output step of outputting the conversion candidates generated by the generation step; An information processing method comprising:

9. an acquisition step of acquiring text data input by a user; a generation step of generating conversion candidates for the text data that satisfy a predetermined notation rule using a trained model that has learned a predetermined notation rule applied within a predetermined group, with the text data acquired by the acquisition step as an input; an output step of outputting the conversion candidates generated by the generation step; An information processing program characterized by causing a computer to execute the above.

Images